When manufacturing composite and laminate products using various fibers, like for instance glass, carbon and aramid fibers as well as flax, hemp, jute, kenaf, basalt and other natural fibers etc. for the manufacture of, for instance, boat, automotive and wind turbine parts, for example, the manufacture starts with the production of appropriate fiber reinforcements like woven or knitted structures, which may have a unidirectional or multi-axial orientation. The structures are, then, placed in a mold used in the manufacture of the intermediate or end product. The mold has, naturally, the shape of the end product meaning that the shape may sometimes be very complicated, and require substantial shaping of the reinforcement when placed in the mold. Normally several layers, up to tens of layers, of reinforcements are stacked one on top of another in the mold and a thermosetting resin like epoxy mixed with hardener or unsaturated polyester resin or vinyl ester is introduced in the mold for forming a fiber-reinforced composite article. Resin may also be thermoplastic like PA (polyamide) or CBT (Cyclic Polybutylene Terephthalate) or alike. Practice has shown that when the end product has to resist high mechanical loads, unidirectional reinforcements, which may be held together in transverse direction by means of stitching, are a preferred choice in its manufacture as they may be oriented efficiently according to loads. Such unidirectional reinforcements are made of rovings or tows, generally called as reinforcing fibers.
The unidirectional reinforcement is normally formed of a single layer of reinforcing rovings. The construction of the reinforcement depends on the target areal weight and the tex number of the rovings. For instance if a high areal weight is desired, a thick roving (for example with E-glass 2400 tex) is used, and where a reinforcement with low areal weight is desired, a thin roving (for example with E-glass 600 tex) is used in its manufacture.
The end product, i.e. the cured laminate structure may be made of a number of such unidirectional reinforcements either by arranging the layers of reinforcements such that, in the end product, the rovings of each layer are parallel or some layers are oriented in other directions according to loads the laminate construction is subjected to or by first manufacturing fabrics of several layers of unidirectional reinforcements so that the rovings of adjacent layers form a certain angle, and thereafter using the fabrics thus formed in the production of the end product. Such fabrics are called biaxial, triaxial, quadaxial etc. fabrics depending on the number of different fiber orientations therein.
A unidirectional reinforcement is inherently unstable in nature as yarns run in one direction only. In order to be able to handle the unidirectional reinforcement, its rovings have to be anchored or bonded to each other in a suitable manner. Prior art knows, in principle, two different mechanical methods for such a purpose.
One method is to secure the rovings by means of stitching (e.g. warp knitting). The stitching yarns form knitting loops, i.e. stitches, which are holding the actual reinforcing rovings, in their place in the reinforcement. The stitches are formed by various knitting elements, e.g. by needles, which penetrate the layer or layers of reinforcing fibers according to the known warp knitting technique. Stitch yarns run lengthwise in the reinforcement and stitches may form several well known patterns like for instance chain or tricot etc. The stitching yarn is typically, but not necessarily, texturized or non-texturized polyester (polyethylene terephthalate) filament yarn having a thickness from about 34 dtex to about 167 dtex and comprising tens of filaments, normally e.g. 24 or 48 filaments.
Stitched reinforcements are well known and they have a few good properties. Firstly, their transverse stability is good if such patterns are used where lengthwise running stitch yarns will be connected with adjacent yarns because the stitching yarns although running mainly longitudinally form such patterns, like tricot, that give the unidirectional rovings the integrity needed for the reinforcement. Secondly, the reinforcement is easy to position in the mold (i.e. make the reinforcement follow the contours of the mold) as the stitched reinforcement is often very flexible if stitching parameters are properly chosen like stitch length, needle gauge and yarn tension, just to name a few as an example.
The use of stitches, however, results in some problems, too. The stitches, for instance, cause regular bends/kinks in the rovings, the bends/kinks deviating the load carrying fibers from their original straight orientation. Stitches also compress locally the rovings causing irregular resin distribution and risk of internal stress concentrations.
Another mechanical method is to use weaving technique to anchor the longitudinal warp yarns with light weight weft yarns in their respective place. As weft yarns both non-coated and hot-melt coated yarns have been used. After heating and cooling the hot melt binder has given the reinforcement considerable stability. Yet the weaving alternative is not any more considered favorable as the reinforcing yarns form kinks when crossing over the weft yarns leading to stress concentrations and lower mechanical properties than knitted versions. The hot melt binder yarns have, when used, been found to create local disturbance in matrix curing and are not favored any more in the trade. Typically, the weft yarns are multifilament yarns that get flat under compression irrespective of their being hot-melt yarns or not.
A further problem concerning reinforcements made by both stitching and especially weaving are the local distortions, i.e. bends or “kinks”, which both the weft yarns and stitches, in a lesser degree, though, create in the otherwise straight reinforcing fiber alignment. The “kinks” lead to stress concentrations and risk of micro-cracking and are considered to be the root cause of reduction in static and dynamic mechanical properties of laminates. This is partly surprising as the stitched reinforcements originally were developed and introduced to the trade as not having “kinks” compared to wovens.
A yet further problem concerning stitched reinforcements is their high flexibility with a clear tendency of easily forming folds or local small curvature bends inside the laminate, especially when the number of reinforcement layers in the laminate is high, which is the case, for instance, in spar cap laminates in wind energy rotor blades. This kind of areas having above described folds or bends become easily rich of resin and may experience excessive high temperatures during resin curing, which easily cause local laminate damage.
A chemical method for bonding the unidirectional rovings together by means of various thermoplastic binders has also been brought to market. Practice has shown that when the rovings are anchored to each other chemically the bonding makes the reinforcement relatively stiff, whereby it is hard to position the reinforcement in the mold, i.e. make the reinforcement follow the contours of the mold, the harder the more complex shapes the mold has. However, choosing carefully both the bonding agent, normally a thermoplastic binder, for instance in powder form, and the amount it is used, the shaping properties (mainly flexibility) of the reinforcement may be brought to an acceptable level. Other problems, in addition to the above discussed handling stiffness, involved in the chemically bonded reinforcements, have been seen in resin permeability and wet-out speed of a stack of reinforcements in a mold.
For solving the problem related to the transverse stability, it has also been suggested that an additional net or layer of, for instance, chopped strand mat or woven net could give a unidirectional reinforcement the transverse stability it requires, but such an additional layer reduces drapeability and increases risk for void formation, weight, thickness and costs of the final product.
The problem relating to resin permeability and to the bonding of the rovings in the reinforcement is caused by the fact that the rovings are locally so tightly against each other in the infusion molding stage, this being emphasized by the compression force of the vacuum infusion process, that, firstly the flow of gas from the open cavities, and secondly, the flow of resin to fill the open cavities in and between the reinforcement layers in the mold is very slow unless any specific measures are taken. As good resin permeability is vital for the practical execution of the molding process it is normally speeded up by utilizing pressure difference when feeding resin in the mold. It is common practice to apply either Vacuum Infusion technology or Resin Transfer Molding (RTM) technology with their numerous variants and flow aid materials like plastic scrims, for distributing the resin all over the reinforcement layers in the mold. However, sometimes despite various measures, like vacuum and/or raised feed pressure, small air cavities tend to remain in the reinforcement reducing significantly the strength properties of the laminate. In view of the above, new ways to improve both the removal of gas from the stack of reinforcements and the permeability of the reinforcement to resin has to be investigated.
Thus both the above discussed basic structures, i.e. mechanically and chemically bonded unidirectional reinforcements have their own problems.
One way to improve the permeability of the reinforcement is to provide the reinforcement with flow passages for resin, the flow passages allowing the resin to flow quickly in the reinforcement. There may be found, in prior art, numerous ways for arranging the resin flow passages in the reinforcements or between the reinforcements in a stack of reinforcements. However, it has been learned that the use of such flow passages is not very efficient, as the vacuum applied in the infusion stage tends to shift or draw rovings from the neighbouring areas or reinforcements and even shift their positions to fill the flow passages/cavities.
EP-A1-1491323 discloses a reinforcement structure comprising unidirectional reinforcement threads and transverse stiffening threads. The stiffening threads are distributed in a spaced manner on a layer of reinforcement threads. The stiffening threads may be of thermoplastic material such that by fusion or softening the stiffening threads fasten to the reinforcement threads and give the reinforcement the transverse stability it needs. For ensuring sufficient capillary draining of injected resin the layer of longitudinal reinforcement threads is provided with longitudinal draining threads, which are, thus, parallel to one another and to the reinforcement threads. The draining threads are arranged in spaced manner in the layer of reinforcement threads. The draining threads may be formed of glass fibres covered with fibres of sufficient capillarity, like for instance cotton fibers or cellulosic fibers, to drain the injected resin. Another option for the draining threads is reinforcement threads on each of which a monofilament is wound around. Thus a spiral flow passage for the resin is formed. Therefore, it is clear that the flow passages in the reinforcement are formed in the longitudinal direction of the reinforcement.
This means, in practice, that the longer the products to be manufactured are the more complex and, at least time consuming, is the impregnation of the end product with resin. In practice, it is impossible to think about impregnating a spar cap of a wind turbine blade having a length of 50 meters or more economically by lengthwise impregnation. Naturally there may be a possibility to arrange resin injections at, for instance, 2 meter intervals over the entire length of a blade, but it is a complicated and time consuming method and, therefore, very expensive.
EP-B1-1667838 discusses the formation of flow passages in a composite fabric formed of a plurality of substantially parallel, coaxially aligned tow groups, each of said tow groups having one or more tows wherein a portion of said tow groups contain two or more tows. The flow of resin along within the fabric is planned to be ensured by arranging spacing between tows in a tow group to be less than the spacing between adjacent tow groups. Thus the spacing between adjacent tow groups should form the required flow passages. Such flow passages should permit resin to flow through the fabric, especially in the direction of the tows, i.e. in the longitudinal direction of the product.
However, as the length of the end product increases it has to be understood that at a certain point the impregnation in longitudinal direction reaches its practical limit, i.e. the so called wet-out distance, whereafter other ways have to be taken into serious consideration. Also, practical experiments have shown that the flow passages will be filled with rovings from nearby areas when vacuum is applied in the infusion stage or the laminate structure becomes corrugated with local kinks in reinforcing rovings reducing mechanical strength.
U.S. Pat. No. 5,484,642 discusses a textile reinforcing material useful for producing composite laminated articles by a general injection-molding technique. The reinforced material, i.e. laminate structure is fabricated by arranging a stack of layers having textile reinforcements in a mold of a shape corresponding to that of the article to be fabricated and, after the mold has been closed, injecting a resin into it. The textile reinforcements may be of woven or non-woven origin including unidirectional slivers. The transverse stability of the reinforcement layers is accomplished by means of weaving, knitting or stitching or by using transverse binding threads or yarns. At least one layer of the stack of textile reinforcements has a structure in which ducts, i.e. flow passages for resin, extend in at least one direction therein to facilitate the flow of the resin during injection. The ducts may be located in longitudinal and/or transverse direction of the material. The main idea behind the above mentioned US patent is to ensure good resin flow properties for the fabric by changing a part of the reinforcing yarns to better withstand compression due to mold closure and vacuum. This is done typically by adding twist to a part of the reinforcing yarns or by twining polyester multifilament yarn around carbon fiber tows. The disadvantage, however, of this concept is that among the normal reinforcing yarns a high number of relatively large yarns are placed that under laminate loading conditions behave quite differently from that of the rest of yarns in the reinforcement. This is mainly due to the often very high twist (260 TPM) that affects the elastic properties of the yarns under loading. Also, the high twist prevents or slows down the resin penetration inside these yarns. This leads to non-homogenous laminate structure where a part of the yarns carries the loads in a different manner. This will finally increase risk of premature laminate failure in static and specifically in dynamic load conditions.
It is worthwhile noting that example 5 of the US-document teaches that the transverse flow passages are formed by arranging weft yarns formed of a 3 K (3000 filaments) carbon thread covered with a polyester thread at 260 turns per meter across the material, whereby spirally advancing flow passages are formed around the covered threads. This could result in good resin flow but 260 TPM is extremely high twist and has a very negative influence on laminate properties according to present state of knowledge.
One well known way of improving the impregnation of resin into a stack of reinforcements is to place in the mold both to the bottom and to the top of the stack a plastic scrim or other flow aid material by means of which the resin spreads quickly to the entire top and bottom surface area of the reinforcement. After impregnation and curing the scrims are removed labour intensively from the laminate. The purpose for the scrim is, naturally, to introduce resin quickly to the entire area of the mold so that the Z-direction impregnation of the resin into the stack of reinforcements could start as soon as possible. However, the thicker the stack is, the slower the stack is to impregnate with resin. For instance in wind turbine blades the cross section of the spar cap is almost a square, whereby, for the resin, the center of the stack is hard to reach.
It is also known that sometimes when using unidirectional reinforcements, especially in woven form, some assisting or additional yarns have been added in transverse direction for improving transverse stability or resin flow properties. Typically the yarns are coated with hot-melt or other thermoplastic material and the yarns are of glass fiber or polyester (for instance, twisted bundles of glass filaments—each bundle having typically 60 or more filaments, each filament having a diameter of 10-15 μm) and in coated form tex number typically 100-200 tex. The thermoplastic coating of the yarns is, after weaving, molten, whereby it flows in the void volumes in connection with both the yarn and the rovings and thus bonds the warp rovings and the weft yarn together. The thermoplastic coating is usually formed of PA (PolyAmide) or EVA (Ethylene-Vinyl Acetate) types of materials, whose melting temperature is lowered by means of waxy substances or by some other appropriate means. Therefore thermoplastic coating is typically conflicting with the infusion resin matrix as the relative amount of binder is locally very high in the immediate vicinity of the reinforcing yarn, causing local weak areas in the laminate. The glass or polyester filaments with glue remain on the rovings transverse thereto and give the reinforcement its transverse handling stability prior to infusion or alike. The resins will not reach the actual fiber surface as the fibers are coated with thermoplastic material.
The use of this kind of assisting or additional yarns in unidirectional reinforcements will increase unnecessarily weight and possibly cause local fiber distortion, which are, by nature, undesirable effects. Furthermore transverse reinforcing fibers, i.e. for instance fibers oriented in 90, 60 or 45 degree direction, may also create micro-cracks when these fibers, normally glass fibers, are broken during the axial loading of the UD-construction from which more severe fatigue cracks destroying the stability of the end product may originate. The reason for the latter problem is the fact that elongation at break of the glass fiber yarn is significantly lower than that of the matrix in transverse direction. And still further the multifilament glass fiber yarns or rovings deform when subjected to vacuum compressing pressure losing their originally round cross section such that their cross section under pressure is oval or even flat (as shown in FIG. 1b). The multifilament yarn form has as a consequence that its individual filaments move sideways leading practically to an oval or flat cross section formation. The yarns coated with thermoplastic material behave similarly as the coating is melting during the heating-pressing stage, leading to flat form where there is a cross-over point.
Firstly, the prior art twisted threads or yarns, i.e. multifilaments used for forming the transverse flow passages have a diameter (before applying compression) of about 0.35-0.45 mm. In the performed tests a laminate was formed by placing a stack of two 1200 g/m2 reinforcement layers having transverse threads of the above mentioned size between the reinforcements in a mold, subjecting the stack to vacuum, performing the infusion with resin, and allowing the laminate to harden. It was learned that the cross section of the multifilament threads was changed to oval or flat while the reinforcement layers were compressed by the applied vacuum in the infusion stage. When comparing the wet-out distance of the reinforcement to that of a reinforcement having no crosswise arranged yarns it was learned that it had not changed or improved at all or the change was, in practice, insignificant. The reason will be discussed in more detail later on.
Naturally, it could also be argued that stitching yarns or corresponding threads running in a direction transverse to the longitudinal direction of the reinforcement rovings could form transverse flow passages for resin. However, it has to be understood that, in addition to the problems involved in the use of stitches discussed already above, the same flattening tendency applies to the stitches and stitching yarns, too.
Reinforcements provided with hot-melt weft yarns have been on the market about 20 years ago, but they have not succeeded in passing strength tests, neither static nor tensile tests. Additionally the plastic formability of such reinforcements was poor. It is, in practice, impossible to use such reinforcements in the production of spar cap laminates for wind turbine blades as the spar caps have double-concave forms to which this kind of reinforcement cannot be bent.
Secondly, reinforcements having transverse glass fiber yarns with a thermoplastic coating have been considered. In such reinforcements the diameter of the coated yarn was of the order of 0.30-0.35 mm, and the diameter, or in fact the Z-direction thickness, of the core yarn of the order of 0.04-0.06 mm when pressed and coating melted or removed. The difference these thermoplastic coated yarns have when compared to non-coated yarns, for example stitching yarns, is that during bonding the yarns to the rovings of the reinforcement, i.e. during the softening/melting of the coating, the yarns change their shape in contact points (compression reduces the Z-direction thickness of the yarns), whereby local flow restrictions are formed. In other words, in points where the coated yarn is not compressed, its diameter remains on the original level, but in points of compression the diameter/thickness is reduced to even below the diameter of the core yarn, i.e. the core of the yarn is flattened by the compression. Another problem involved in the use of coated yarns is that the yarn is stiff and relatively thick, whereby it makes the direction of the rovings deviate sharply locally from their straight direction, i.e. forces the rovings to bend and form kinks with problems discussed already earlier as well as later on in this paragraph. A yet further problem with the coated yarns is the coating polymer itself, as it is normally not compatible with the resin and thus contaminates the laminate and thus creates weak spots in the reinforcement. Now a laminate was formed of reinforcement layers each bonded by the transverse coated glass fiber yarns to give stability. It was learned that the wet-out distance of the stack of reinforcements was mostly acceptable. But when the laminate, thus having transverse glass fiber yarns with a diameter or thickness changing between about 0.35 and about 0.04 mm, was subjected to fatigue testing, it was learned that soon after the start of the tensile-tensile fatigue testing micro cracking of the laminate was observed. When examining the laminate and especially the micro-cracks in detail it was learned that the micro-cracks were found at the junctions of the reinforcement rovings and the coated transverse yarns. A clear doubt was that the reason for the micro-cracking was the local large diameter of the thread creating bends or kinks in the rovings. Additionally, the hot melt yarns, i.e. also the core yarns, when heated, are compressible, whereby local flattened areas are created, which reduce the cross section of flow passages and thus hamper resin flow at the infusion stage.
Thus both ways of holding the unidirectional reinforcement in shape or undistorted in transverse direction have their own drawbacks. However, as the stitching has a tension-related problem, which is caused by the basic feature of stitching, i.e. the tightening of a stitch round one or more rovings, it is, in practice, extremely difficult to get rid of this basic feature. Thus, it appears that the bonding of the rovings to hold the shape by means of a thermoplastic binder is probably the way the reinforcements will be made in the future, specifically if fatigue properties need to be optimized. Therefore, new ways to improve the permeability of the reinforcement to resin have been investigated.
As the starting point for the further development of an optimal unidirectional reinforcement is a reinforcement in which the problems relating to handling stability and resin permeability have been taken care of without sacrificing the strength and especially the fatigue properties. For ensuring maximal strength and fatigue properties for the laminate manufactured of this reinforcement, there should not be any yarns/fibers stretching crosswise over the reinforcement for giving the material stability in transverse direction. However, it is very difficult to handle the material during the subsequent work phases. The often heavy reinforcements are laid in the mould, for instance during the wind turbine blade production phases, often several, sometimes even 50-60 layers one on top of another, the reinforcements having a length of several meters, often up to 50-60 meters, sometimes even more. The transverse stability is needed so that the operator stacking the reinforcements in the mold in the blade production is able to adjust the exact position of each reinforcement in transverse direction. This is not possible without sufficient transversal pulling strength. If the flow passages extend in longitudinal direction continuously from one end of reinforcement to the opposite end thereof, like the EP-B1-1667838 teaches, the channels form weak lines along which the reinforcement is easily torn into two or more parts, i.e. into longitudinal stripes.
The practice has shown that the present day unidirectional reinforcements have several problem areas, like for instance:                the unidirectional reinforcement has in its basic, powder bonded form, extremely limited permeability to resin in general and especially in transverse direction, therefore the production of lengthy objects is extremely challenging,        if the permeability is improved by longitudinal flow passages of the prior art teachings for the resin, the reinforcement loses its transverse stability,        stitching and the use of transverse coated yarns creates kinks in the rovings resulting in risk of micro cracks, and reduced strength properties,        stitched reinforcements have a high risk to form folds or small curvature bends inside thick sections of laminate when laying them into a mold. Folds may cause local high exotherm regions with serious laminate damage.        transverse thick twisted multifilaments aimed at providing flow passages for resin and being arranged between and transverse to the unidirectional reinforcements force the rovings too far from another and result in high risk of micro cracks, and vacuum pressure compresses the multifilaments resulting in much less effective diameter, i.e. Z-direction thickness, flattened transverse yarns prevent resin from flowing in these flow passages,        gas bubbles or dry regions between the filaments of UD rovings remain easily in the stack of reinforcement and cannot be removed even in vacuum infusion, whereby they may considerably reduce the strength of the end product even further, and        thermoplastic coating and hot-melt glue conflicts with the resins used in the infusion at least locally.        